Kemira, CuspAI explore 300-trillion PFAS

Kemira and CuspAI said on May 21 that they had used generative artificial intelligence to design new materials aimed at removing PFAS from drinking and process water at trace concentrations. The companies said the project explored about 300 trillion possible material structures, generated more than 5,000 novel designs and narrowed those to roughly 20 priority candidates for further development. Kemira described the work as the first commercial partnership to apply generative AI end-to-end to designing new materials for PFAS remediation. That claim matters because PFAS — short for per- and polyfluoroalkyl substances — are a broad class of chemicals often called “forever chemicals” because they are highly persistent. Kemira said the initial targets in the project were three PFAS molecules: GenX, PFBS and PFOS. The company said the resulting candidates were evaluated for industrial performance criteria including stability, sustainability and manufacturability. (kemira.com) ### Where did the 300-trillion number come from? Kemira said the materials discovery program searched a design space of approximately 300 trillion possible material structures. The company said CuspAI’s platform designed entirely new structures from scratch rather than only screening known materials, and that the program reached its current stage in six months. (kemira.com) CuspAI was announced as Kemira’s partner on July 9, 2025, when the companies said they would use AI to accelerate discovery and optimization of new materials. In that earlier announcement, Kemira said materials discovery that can take up to a decade could be compressed to as little as six months with AI, with PFAS removal chosen as the first project. (kemira.com) ### What exactly did the companies say they found? Kemira said the project delivered more than 5,000 novel material designs with full property data across the three priority PFAS molecules. The shortlist was then reduced to about 20 selected priority candidates that are now moving into the next phase of development. (kemira.com) The announcement did not say that the materials had been validated in commercial water-treatment systems or approved by regulators. Instead, Kemira said the candidates are advancing to further development and testing, which means the current result is a narrowed set of AI-designed materials rather than a deployed product. That is an inference from the company’s description of the stage reached. (kemira.com) ### Why were Kemira and CuspAI working on PFAS? Kemira said PFAS and other micropollutants are becoming more relevant in water treatment because of consumer-health and environmental concerns and because of regulation. The company has also been building out its water-treatment offering in activated carbon, which it has described as a common technology for removing micropollutants including PFAS. (kemira.com) Kemira’s 2025 annual materials said the CuspAI partnership was intended to accelerate materials innovation from years to months, starting with PFAS removal. Those disclosures tie the project to Kemira’s broader water-solutions business rather than a one-off research exercise. ### What did the companies say about the next phase? Dr. Chad Edwards, CuspAI’s CEO and co-founder, said on May 21 that the partnership had delivered in six months and was “advancing to its next phase.” Sampo Lahtinen, Kemira’s executive vice president for research and innovation, said the teams had evaluated candidates against “real industrial requirements,” while Kemira CEO Antti Salminen said the work showed what AI-driven materials design looks like when paired with industrial use cases. (kemira.com 1) (kemira.com 2) Kemira said the next step is further development and testing of the roughly 20 selected candidates. The company has not yet given a date for pilot results or commercialization. (kemira.com)